Detailseite
Projekt Druckansicht

Counting without a cortex - Neuronal correlates of quantity representations in the corvid brain

Fachliche Zuordnung Kognitive, systemische und Verhaltensneurobiologie
Förderung Förderung von 2012 bis 2017
Projektkennung Deutsche Forschungsgemeinschaft (DFG) - Projektnummer 212142797
 
Erstellungsjahr 2018

Zusammenfassung der Projektergebnisse

Cognitively-advanced corvids provide a “window of opportunity” to decipher the neurophysiological mechanisms of numerical representations based on different endbrain structures. With this funding, we were able to establish electrophysiological experiments in behaving carrion crows (Corvus corone) aimed at elucidating the workings of the NCL in number processing. The derived data provide novel insight into the role of an important corvid association area in complex numerical cognition. Classifying items according to their number is a particular abstract category. It allows animals to abstract from the sensory attributes of elements to arrive at the quantitative category “numerosity”. Ever since Koehler and his co-workers explored the numerical capabilities of birds, corvids have been known to excel relative to other birds. In our tasks that required absolute numerosity discriminations, carrion crows were able to discriminate the specific number of items in small and large visual item sets. They managed numerosity discriminations irrespective of the physical appearance of the array. For instance, when four dots were displayed in a square-like arrangement or as a straight line, both were judged as instances of numerosity four. Crows grasp the concept of numerical quantity. In subsequent recordings in discriminating crows, we found many neurons in the NCL that were tuned to individual numerosities ranging from 1 to 30 items. Such neurons showed the largest discharge rates for their “preferred numerosity”. The characteristics of both the neuronal and the behavioral tuning functions showed close correspondence. Moreover, numerosity selectivity was impaired during error trials. Both findings argue that the neuronal discharge of NCL neurons is the neuronal basis for the crows’ performance. Numerosityselective neurons in the crow NCL is not caused by training, because we also found such neurons in numerically-naïve crows that have never been trained to discriminate set sizes. As an interesting extension of the originally proposed research plan, we also recorded from the crow hippocampus while the corvids performed a delayed match-to-numerosity task. The lack of numerosity-selective neurons in the hippocampus suggests two major conclusions: First, numerosity-selectivity is processed in special brain networks of the avian brain, such as the NCL. Compared to the NCL, the hippocampus serves as a control recording area that obviously does not belong to the avain number network. Second, the avian hippocampus does not seem to be involved in tasks that do require non-spatial working memory. It seems that the avian and mammalian hippocampi are not only dissociated in terms of anatomical connectivity, but also in terms of function. To sum up, we managed to answer the main questions of the project: Numerosity-selective neurons exist in the NCL of numerically-trained and numerically-naïve crows. The observed code in the crow brain is surprisingly similar to the one observed previously in nonhuman primates. This way of coding numerical information seems to have evolved based on convergent evolution, probably as a superior solution to a common computational problem. We hope our studies underscore the value of a comparative systems neuroscience approach, which we believe is indispensable for deciphering evolutionary stable neuronal mechanisms and codes.

Projektbezogene Publikationen (Auswahl)

 
 

Zusatzinformationen

Textvergrößerung und Kontrastanpassung